DimensionalMechanics develops off-the-shelf pattern recognition and analysis tools so companies don’t need to build their own artificial intelligence software.
What: DimensionalMechanics, a builder of artificial-intelligence tools
Who: Rajeev Dutt, president and CEO
The team: Dutt, who has a background in theoretical physics and has done software-development work at Hewlett-Packard, Microsoft and Intel, oversees an 11-person team that includes software developers, mathematicians, data-science experts and a psychologist.
“What’s nice about the Seattle area is there’s a growing job force, and talent in educated people who are skilled at doing these kinds of things,” Dutt said.
Not quite robots: Artificial intelligence is a pretty broad term (and often misused, industry insiders say). When Dimensional execs say A.I., they don’t mean a fully conversant robot that can also fold your laundry. Instead, they’re building a software platform that provides off-the-shelf pattern recognition and analysis tools. The idea is to enable businesses to take advantage of advances in smarter computing without employing legions of mathematicians and software developers.
Media focus: An early customer target for Dimensional is broadcast-media companies, Dutt says. In the fierce competition for attention on the internet, media groups could be helped by software that can spot patterns and tell them what about a particular video or headline was a hit with users, or indicate where they can improve. Dimensional also sees potential applications in medicine, defense, video gaming, design, and other data-intensive industries.
First product: An early version of Dimensional’s software platform, a trio of services on a platform dubbed NeoPulse, was released this month, and Dutt says several customers are trying out the company’s tools.
Built on the local cloud: Dimensional’s software, which can require a lot of computing horsepower for complex tasks, would have required a massive investment in server hardware a decade ago. But in the era of on-demand computing power delivered over the internet, Dimensional uses Amazon.com’s AWS and Microsoft’s Azure to do the heavy lifting. The company relies on AWS data centers for tasks requiring its high-performance computing prowess, and on Azure for tools that need to play well with Microsoft’s other products.
The money: Dimensional raised $4.7 million in venture capital earlier this year, and Dutt says it’s preparing for another round, likely to close in 2017.
The machines are learning: “There are many different ways to teach a machine how to learn something,” Dutt says. “The problem is that not every method is suitable for every problem. What we’re trying to do is [build a tool for] the smaller companies that might not have the skillsets” to wade into A.I. on their own, he said.